observation · Misc · computed
Root cause of Scotchn's latching SCR bench: he used the 8XA SCR-as-gate inside the 20YI topology. 8XA gating works only into a diode/choke/cell load, not into a transformer primary
Scotchn, Discord #neuralstan20-testing 2026-11-04: "On the bench it didn't work because I'd confused the 8XA and GAS Resonant Cavity schematics and had the 8XA SCR gate in the GAS Resonant Cavity topology!" Lesson (my analysis): in 8XA the SCR is the gate itself and feeds blocking diode > bifilar chokes > cell; the series diode plus capacitive cell lets the current die (rail valley commutation, #3363). In 20YI the SCRs only select amplitude, and the gating is done by a separate chopper Q9 into an inductive primary; Q9's off-time is what commutates the SCRs. Using 8XA's SCR gating into a 0.83 H primary gives continuous conduction through the bridge freewheel, so it latches (#3518). Closes #3517-#3520. The two sheets are not interchangeable at the gate stage.
Basis
- Recorded
- Published
- 4 Nov 2026
- Stanbot
- v3
- Source Ref
- Scotchn Discord 2026-11-04; notebook #3363, #3514-#3520
- Notebook Id
- 3521
- Device Slug Unmapped
- hgg
20YI 8XA SCR root-cause scotchn